A Collaborative Filtering Recommendation Technology Based On Clustering And User Preference

With the rapid development of the Internet and e-commerce,bring people living convenience and rich choice,but at the same time also produced information overload and other issues.How to quickly find out our needs in the huge commodity information,has become an urgent problem to be solved.In order to solve this problem,recommendation technology came to our life.Collaborative filtering algorithm can predict their possible interest based on the known preferences of users,is now the most successful and widely used recommended technology.However,the traditional collaborative filtering recommendation algorithm is limited by sparseness of data,resulting in poor recommendation.Aiming at the problems above,The main work of this paper is as follows:(1)A score matrix model based on user preference is constructed.By combining the original scoring matrix and the project type information,this paper constructs the user interest preference matrix.Based on the user interest preference matrix,it can effectively find the neighborhood user set with similar user's interest preference to the target user.(2)A weighted Slope One algorithm based on the similarity between projects is proposed.Aimming at the sparseness of the original "User-Item" scoring matrix,in this paper,the similarity between projects is integrated into the Weight Slope One algorithm,and an evaluation model is proposed,which can effectively fill the original scoring matrix to form a more dense scoring matrix.Based on the filled-scoring matrix,we can find the target users nearest neighbor users more accurately,therefor could further improve the prediction accuracy of the proposed algorithm.(3)This paper proposes a collaborative filtering recommendation technology based on clustering and user preference.Firstly,the evaluation model of weighted Slope One algorithm based on the similarity between projects is used to fill the original "User-Item" scoring matrix,and then the K-Means algorithm is used to cluster the users based on the user interest preference matrix.Finally,the User-Based collaborative filtering recommendation is implemented in each user cluster combine with the padded matrix.(4)Experimental results and analysis.Based on the MovieLens data set,use experiment to verify collaborative filtering recommendation algorithm proposed in this paper,the experimental results show that the algorithm proposed in this paper can effectively alleviate the sparseness problem of traditional collaborative filtering algorithm,and improve the recommended quality of the algorithm.